Comment on "Under-reported data analysis with INAR-hidden Markov chains"

Johannes Bracher
DOI: https://doi.org/10.48550/arXiv.1812.06688
2019-02-28
Abstract:In Fernandez-Fontelo et al (Statis. Med. 2016, DOI <a class="link-https link-external" data-doi="10.1002/sim.7026" href="https://doi.org/10.1002/sim.7026" rel="external noopener nofollow">https://doi.org/10.1002/sim.7026</a>) hidden integer-valued autoregressive (INAR) processes are used to estimate reporting probabilities for various diseases. In this comment it is demonstrated that the Poisson INAR(1) model with time-homogeneous underreporting can be expressed equivalently as a completely observed INAR(inf) model with a geometric lag structure. This implies that estimated reporting probabilities depend on the assumed lag structure of the latent process.
Methodology
What problem does this paper attempt to address?